On May 22, 2008, at 10:35 AM, John C Klensin wrote:
--On Thursday, 22 May, 2008 10:15 -0400 Ed Juskevicius
Every so often someone suggests RFCs are not first class
documents and hence not comparable to, say, "real"
standards documents. Getting traditional identifiers attached
to them might squelch some of this nonsense.
I have the impression that we would be pioneering the use of
an ISSN to identify a standards' series, if we choose to do
this. The "real" standards from other organizations seem to
be identified with individual ISBNs.
If all RFCs were standards, this would be a good argument.
However, the RFC series contains Standards, various types of
substandards for which those other bodies either have no
equivalents or publish differently, experimental protocol
specifications, BCP statements, and an assortment of
informational documents. What an ISSN identifies is the
series, not the individual documents, and that series is _not_ a
My impression is that this type of application is not
particularly novel. More on that next week.
As I indicated in my note to Melissa, having an ISSN for the
series would not prevent obtaining ISBNs and/or DOIs for
selected individual documents, so those ideas are really
completely separate questions.
Would the purveyors of nonsense be squelched by an ISSN, or
emboldened? Some might cite our decision as yet another
example of the IETF doing something different and
Very unlikely. At worst, we would be "accused" of illustrating
ways in which an existing standard mechanism can be carried
forward in interesting ways into the modern Internet age. On
the other hand, if we treat RFCs as basically paper (and
page-format) publications that are freely available online as I
suggested in an earlier note, this becomes that most routine of
Here is a concrete suggestion.
We (for some definition of we) have the Internet Journal, which is
Publish a "Supplement of the Internet Journal," in paper, or on line,
- physically published 3 times a year
- has all of the RFC's published since then
- includes the level 1 RFC errata as available
- includes other notes like RFC's that have been made obsolete, etc.
- charge it to cover costs at least (say, $ 500 / year for a
This would be picked up by at least some libraries, and would solve
the "on-line is ephemera" problem.
Marshall, to your point:
It is easy to find RFC's now, but it may not be in a century.
This may seem silly, but I think that RFCs will still
have relevance in a century and, having experience
searching for 100+ year old astronomical publications
and data, in my opinion, RFC's need to be cataloged in
Libraries have running code for the maintenance of
intellectual property over centuries; the IETF does not.
I agree with you 100%. I think this is indeed a tangible and
Indeed. And libraries, especially the subset of libraries that
have national archival responsibilities, do pay attention to
IETF mailing list